Face recognition based on principal component analysis is a heavily researched topic in computer vision. The ear has been proposed as a biometric, with claimed advantages over the face. We have applied the PCA approach to images of the face and ear using the same set of subjects. Testing wass done with three different gallery/probe combinations. For faces we have: 1) probes of same day but different expression, 2) probes of a different day but similar expression, and 3) probes of different day and different expression. Analogously, for ears, we have: 1) probes of same day but other ear, 2) probes of a different day but same ear, and 3) probes of different day and other ear. Results indicate that the face provides a more reliable biometric than the ear.